The heterogeneous selection landscape of genome evolution in prokaryotes

The heterogeneous selection landscape of genome evolution in prokaryotes | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results The heterogeneous selection landscape of genome evolution in prokaryotes Roman Kogay , Svetlana Karamycheva , View ORCID Profile Nash D. Rochman , Yuri I. Wolf , Eugene V. Koonin doi: https://doi.org/10.1101/2025.11.26.690804 Roman Kogay 1 Computational Biology Branch, Division of Intramural Research, National Library of Medicine, National Institutes of Health , Bethesda, MD 20894, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Svetlana Karamycheva 1 Computational Biology Branch, Division of Intramural Research, National Library of Medicine, National Institutes of Health , Bethesda, MD 20894, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Nash D. Rochman 1 Computational Biology Branch, Division of Intramural Research, National Library of Medicine, National Institutes of Health , Bethesda, MD 20894, USA 2 City University of New York Graduate School of Public Health and Health Policy, Department of Epidemiology and Biostatistics , New York, NY, United States 3 Institute for Implementation Science in Population Health, City University of New York School of Public Health , New York, NY Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Nash D. Rochman Yuri I. Wolf 1 Computational Biology Branch, Division of Intramural Research, National Library of Medicine, National Institutes of Health , Bethesda, MD 20894, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Eugene V. Koonin 1 Computational Biology Branch, Division of Intramural Research, National Library of Medicine, National Institutes of Health , Bethesda, MD 20894, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: koonin{at}ncbi.nlm.nih.gov Abstract Full Text Info/History Metrics Supplementary material Data/Code Preview PDF Abstract Evolution of prokaryote genomes appears to be defined by the interplay of selection for genome streamlining, deletion bias and selection for functional diversification. The previously observed overall positive correlation between the strength of selection, measured as the ratio of non-synonymous to synonymous nucleotide substitutions ( dN/dS ), points to diversification as the primary factor of prokaryote genome evolution. Here, we investigated the interplay between genome size and selection pressure by analyzing an expanded collection of closely related prokaryotic genomes, evaluating genome-wide selection by measuring dN/dS by using an accurate, phylogeny-based method and decomposing the resulting values into lineage-specific and gene-specific components. These analyses reveal a pronounced heterogeneity in the relationship between genome size and the strength of selection across the diversity of prokaryotes. Most bacteria display a positive correlation consistent with selection for diversification, whereas all analyzed archaeal lineages show strong negative correlation which is the signature of streamlining. These findings indicate that the selection regimes broadly vary across the diversity of prokaryotes rather than following a single, universal pattern. Genome streamlining, selection for functional diversity and drift in small populations are all important factors of evolution, their relative contributions depending on the population genetics and ecology of a given lineage. Introduction Evolution of genome complexity that is reflected in the number of distinct genes is thought to be shaped primarily by the product of selection and effective population size ( Ne ) 1 . Within this population-genetic framework and considering the universally observed excess of deletions over insertions (deletion bias) in the evolution of genomes 2 , 3 , prokaryotic evolution is often viewed as being dominated by selection for genome streamlining 1 , 4 , 5 . Indeed, some of the most abundant (very large N e ) free-living bacteria, notably those adapted to oligotrophic marine conditions, such as members of the SAR11 clade, have the smallest genomes among free-living bacteria and evolve under strong purifying selection, in the streamlining regime 4 , 6 . However, streamlining Is not the only major factor in the evolution of prokaryote genomes that exhibit broad diversity in gene number and composition, ranging over two orders of magnitude in size, from tiny, highly reduced genomes of endo- and ecto-symbionts 7 , some of which encompass less than 200 genes, to gene-rich genomes of many free-living bacteria, some containing more than 11,000 genes 8 – 11 . With the sequencing of first genomes of obligate endosymbionts, one of the most striking observations was the pervasive reduction in their genome sizes 12 , 13 . These drastically eroded genomes have lost a large fraction of metabolic and regulatory genes, retaining only core functions required for translation, replication and host dependency 14 . Such extreme genome reduction appears to be a combined effect of genetic drift and a pronounced deletion bias 3 . Small N e that is typical among bacteria with host-restricted lifestyles substantially limits the efficiency of selection, facilitating accumulation of deleterious deletions, which eventually results in a gradual genome contraction and functional simplification 15 . Thus, genome contraction in prokaryotes can occur both under strong selection for streamlining and via a deletion ratchet under weak selection. In contrast, the selection for diversification favors the expansion of genetic and regulatory repertoire, promoting metabolic versatility, and regulatory complexity 16 . Generally, the strength of selection in prokaryotes is proxied by the ratio of non-synonymous to synonymous mutations ( dN/dS ) in protein-coding genes which reflects the strength of selection on the protein level 17 , which decreases with increasing N e 18 . Previous studies have demonstrated that dN/dS in prokaryotes generally tends to negatively correlate with the genome size (total number of genes) 16 , 19 – 21 , being partially driven by underlying phylogeny 22 , which appears to be poorly compatible with the streamlining regime of evolution. Therefore, it has been proposed and encapsulated in a mathematical model that positive selection for functional diversification that results in acquisition of multiple genes, primarily, via horizontal gene transfer (HGT), overrides both purifying selection for streamlining and the deletion ratchet in prokaryote evolution 21 . However, due to the methodological limitations, these studies primarily estimated dN/dS from pairwise genome comparisons, summarizing the results with median or mean values. This approach obscures phylogenetic nonindependence within lineages and inadequately accounts for shallower and deeper divergences within the group. Furthermore, previous analyses were primarily limited to conserved housekeeping genes that can potentially misrepresent genome-wide selection landscape if core genes experience different constraints than accessory genes. Most importantly, averaging of the correlations due to limited genomic data obscured potential heterogeneity of the selection landscapes across the diversity of bacteria and archaea. Here, we sought to investigate how genome size relates to selection pressure using an expanded collection of closely related prokaryotic genomes from the latest version of the Alignable Tight Genomic Cluster (ATGC) database, containing over 22,000 genomes 23 . We evaluated genome-wide selection by measuring dN/dS for various orthologous gene families using a phylogeny-aware approach and partitioned those values into lineage-specific and gene-specific components. The results of these analyses reveal a notable heterogeneity in the relationship between genome size and the strength of selection across the diversity of prokaryotes. Strikingly, most bacterial lineages display a positive correlation consistent with selection for diversification, whereas all analyzed archaeal lineages show strong negative correlation that appears to be the signature of streamlining. We demonstrate that the observed trends persist after controlling for sampling bias and are robust across various functional categories of genes. These findings indicate that the selection regimes vary broadly across the diversity of prokaryotes rather than following a single, universal pattern. Genome streamlining, selection for functional diversity and drift in small populations are all important factors of evolution, their relative contributions depending on the population genetics and ecology of a given lineage. Results Lineage-specific and gene-specific selection in prokaryotes Accurate estimation of dN/dS values requires alignments that include sufficient numbers of both synonymous and nonsynonymous mutations, while avoiding saturation that can obscure evolutionary rates 24 – 26 . To achieve this balance, we initiated our analyses using the latest version of the ATGC database, in which each ATGC consisted of at least four genomes that shared ancestral gene synteny and stay below the saturation limit at synonymous sites. To further ensure a robust signal, we retained only those ATGC COGs that encompassed at least 32 mutations across the alignment and only those ATGCs where at least 70% of orthologs satisfied this condition. Under these criteria, the final dataset consisted of 791 ATGCs ( Figure S1 ), including 19 archaeal and 772 bacterial ones. Genome sizes within this collection of ATGCs spanned more than two orders of magnitude: the smallest genomes belonged to ATGC1045 containing members of the genus Nasuia (median length of 0.11 Mb), whereas the largest genomes were found among Streptomyces species in ATGC0548 (median length of 12.02 Mb) ( Table S1) . To validate the reliability and accuracy of dN/dS estimations for gene alignments using IQ-Tree 27 , we conducted simulations that mimicked empirical patterns of sequence evolution in representative bacterial lineages. Specifically, we used the phylogeny of ATGC0001 (comprising Escherichia , Salmonella and Citrobacter ) as a guide tree and generated 2,724 artificial codon alignments. For each simulated alignment, we used the observed substitution patterns ( dN/dS and transition/transversion ratio), alignment length, and branch lengths estimated from the corresponding clusters of orthologous genes (COGs) 28 . The resulting simulated alignments were then reanalyzed with IQ-Tree using the same mode selection procedure as applied to empirical alignments. The results demonstrate that IQ-Tree accurately recovered both dN/dS (Spearman’s R = 0.96, asymptotic p-value < 1 ∗ 10 −16 ; Figure S2A ) and transition/transversion ratios (Spearman’s R = 0.90, asymptotic p-value < 1 ∗ 10 −16 ; Figure S2B ), confirming its suitability for large-scale analyses of the selection regime in the evolution of bacterial and archaeal genomes. To estimate the genome-wide selective constraints across the entire set of genes, including those that are not represented in (nearly) all ATGCs and to characterize the selection regimes more thoroughly, we decomposed dN/dS values across all COGs into components reflecting lineage (ATGC)-specific constraints, gene (COG)-specific constraints, and residuals. First, each ATGC COG was linked to the prokaryote-wide COGs 28 (pCOGs here) by sequence similarity; when several paralogous ATGC COGs mapped to the same pCOG, the ATGC COG with the largest number of genes was chosen as the representative. For each , where L corresponds to lineage (ATGC), and G corresponds to gene (pCOG), we modeled the data as: where F L is the lineage-specific component, R G is the gene-specific component, and e G , L is the residual term (since r L , G is in log scale, F L and R G are logarithms of the multiplicative factors). The parameters F L and R G were estimated iteratively by minimizing the sum of squared residuals: At each iteration, the mean of the residuals was calculated across all COGs for each ATGC to update F L , and then, across all ATGCs for each COG to update R G . These steps were repeated until convergence (change in E between iterations lower than 10 −6 ) or a maximum of 1,000 iterations. When estimating the strength of selection, it is generally assumed that selective forces act relatively uniformly across the evolutionary unit under study. To assess whether this assumption held for the ATGCs, we initially identified 30 ATGCs for which the phylogenetic tree could be partitioned into two subtrees, each containing at least 20 genomes with sufficient phylogenetic depth (see Methods). For each of these ‘splittable’ ATGCs, we calculated the dN/dS values for all COGs across the ATGC as a whole and separately for each of the two subtrees. In some ATGCs, selection patterns appeared visually uniform between the two subtrees, whereas others exhibited notable differences ( Figure S3 ). To test whether these differences could arise from stochastic variation, we performed a bootstrap analysis. For each splittable ATGC, we randomly selected 100 COGs and calculated dN/dS values for 100 bootstrap replicates, measuring the proportion of samples that exceeded the value observed for the corresponding COG in the full ATGC ( Figure S4 ). Although some variation in selection between subtrees was evident, in most ATGCs, only a small fraction of genes showed significant interclade differences within ATGCs. The estimated ATGC-specific (lineage-specific) constraints suggest that only 6 of the 30 splittable ATGCs display moderate to strong heterogeneity of the selection pressures (>40% difference between subtrees; Table S2 ). Thus, most ATGCs exhibited relatively uniform selective regimes, validating their suitability as coherent units for comparative analyses of selection. Different selection regimes across the diversity of prokaryotes To quantify the relationship between selection strength and genome size in bacteria and archaea, we aggregated dN/dS values from all COGs for 791 retained ATGCs, computed their ATGC-specific selective constraints, and used each ATGC’s median genome size as a representative ( Figure S5 ). In this analysis, the ATGC-specific constraint was used as the proxy for selection strength in order to isolate the signal of potentially different selection regimes across bacterial and archaeal lineages. For the entire set of the ATGCs, a significant positive correlation between the genome size and the ATGC-specific constraint was observed, in agreement with previous studies 16 , 19 – 21 and consistent with selection for diversification (Spearman’s R = 0.26, 95% CI = 0.19 – 0.34; asymptotic p-value = 5 ∗ 10 −14 ) ( Figure 1A ). Furthermore, also in agreement with the previous study 22 , the ATGCs with intermediate genome sizes showed a wide spread in selective constraint values, whereas the largest genomes consistently exhibited moderate constraint. Download figure Open in new tab Figure 1. Relationship between genome size and ATGC-specific constraints across prokaryotes. (A) Scatterplot of genome size versus ATGC-specific constraint, which represents lineage-level selective pressure estimated from the decomposition of dN/dS values across orthologous genes. Each point corresponds to a bacterial or archaeal ATGC, with the genome size represented by the median value for that ATGC. (B) Distribution of Spearman’s correlation coefficients between genome size and ATGC-specific constraint across the prokaryotic phylogeny. Major bacterial phyla containing more than 15 ATGCs, as well as Archaea, are highlighted with distinct colors. Projecting bars indicate median genome sizes for individual ATGCs. Internal nodes are colored according to the correlation among their descendant lineages; only nodes with more than 15 descendants are directly colored, whereas others inherit the color of their parent node. Mapping the Spearman’s R values between genome size and ATGC-specific selective constraint onto the phylogeny revealed pronounced heterogeneity across the prokaryotic tree, with the most pronounced contrast observed between the two prokaryotic domains of life, Bacteria and Archaea ( Figure 1B ). Bacteria, which dominate the ATGC dataset, showed a significant positive correlation between genome size and selective constraints (Spearman’s R = 0.28, 95% CI = 0.20 – 0.35; asymptotic p-value = 6 ∗ 10 −15 ) ( Figure 2A ), implying selection for diversification. In contrast, Archaea exhibited a strong negative correlation between genome size and ATGC-specific selective constraints (Spearman’s R = -0.82, 95% CI: -0.93 – -0.64; asymptotic p-value = 1.8 ∗ 10 −5 ) ( Figure 2B ). Similar to a previous report showing a positive correlation between dN/dS and genome size in Archaea 29 , our results support the notion that larger archaeal genomes evolve under weaker purifying selection, which is consistent with the streamlining selection regime. To test whether the strong negative correlation between genome size and selection strength observed for Archaea could reflect sampling bias and imbalance in the analyzed datasets, we resampled bacterial ATGCs 10,000 times (n=19) and calculated the correlation coefficient for each draw. None of the samples had a R value less than -0.82, with the minimum observed value of -0.69, indicating that the negative correlation between genome size and the strength of selection in archaea could not be explained by sampling effects ( Figure S6 ). Download figure Open in new tab Figure 2. Contrasting relationships between genome size and ATGC-specific constraints in Bacteria (A) and Archaea (B). Left panels show scatterplots illustrating the relationship between ATGC-specific constraint and genome size. Right panels display the distributions of Spearman’s correlation coefficients obtained from 1,000 bootstrap replicates, which were used to calculate confidence intervals. Although the overall correlation between genome size and selection strength among bacteria supports the selection for diversification regime, a more granular examination reveals pronounced heterogeneity across bacterial taxa. At the phylum level, strong positive correlation was observed for the vast phylum Pseudomonadota (Spearman’s R = 0.38, 95% CI = 0.27–0.49; asymptotic p-value = 1.1 ∗ 10 −12 and for Bacteroidota (Spearman’s R = 0.35, 95% CI = 0.09-0.59; asymptotic p-value = 0.006), whereas for the other phyla, no significant relationship was detected ( Table S3) . However, a finer taxonomic resolution reveals additional pockets of the diversification regime, such as orders Mycobacteriales (Spearman’s R = 0.41, 95% CI = 0.09 – 0.66; asymptotic p-value = 0.007) and Lactobacillales (Spearman’s R = 0.31, 95% CI = -0.01 – 0.56; asymptotic p-value = 0.031); these signals become masked when aggregated with other clades from the same phyla ( Figure 1B , Table S3 ). Negative correlation implying streamlining is rare among the analyzed Bacteria , prominently observed only in the order Bacillales (Spearman’s R = -0.33, 95% CI = -0.52 – 0.12; asymptotic p-value = 0.009). Consistency of selection regimes across functional groups of prokaryotic genes We further tested whether selection regimes varied among genes with different functions by grouping COGs into four classes: i) information storage and processing, ii) cellular processes and signaling, iii) metabolism, and iv) poorly characterized genes, and computing ATGC-specific selective constraints for each class independently. All four classes showed comparable overall positive correlations with genome size (Spearman’s R = 0.23 – 0.28, asymptotic p-values < 2.7 ∗ 10 −11 ) ( Figure 3 ), and near perfect concordance with the genome-wide ATGC-specific constraints (Spearman’s R = 0.97 – 0.99, asymptotic p-values < 1 ∗ 10 −16 ) ( Figure S7 ). Thus, selective pressures appear to be broadly shared across functional classes of genes. At finer resolution, constraints estimated for individual COG categories also closely matched genome-wide constraints and one another, reinforcing the conclusion of a uniform selective regime for most genes ( Figure S8 ). A notable exception is the X category (mobilome), which showed a much weaker correlation with other COG categories ( Figure S8 ). This deviation suggests that mobilome genes are substantially affected by other evolutionary forces, such as spread via HGT, making them less tightly coupled to the genome-wide selection landscape. Download figure Open in new tab Figure 3. Relationship between genome size and ATGC-specific constraints across major functional categories of genes (COGs). ATGC-specific constraint represents lineage-level selective pressure estimated from the decomposition of dN/dS values within each functional subset. Only ATGCs with dN/dS estimates for at least 10 COGs were included in the analysis. Association between gene flux and selective constraints To evaluate the association between the selection regimes and the overall genome dynamics, we quantified the relationship between gene flux rates and ATGC-specific selective constraints. Specifically, we tested whether the rates of gene gain, gene loss, and overall flux (gain/loss ratio) correlated with the strength of selection across ATGCs. No significant correlation was identified between the selective constraints and gene gain (Spearman’s R= 0.01, 95% CI: -0.06 – 0.09; asymptotic p-value = 0.78) or gene loss rate (Spearman’s R = -0.02, 95% CI: -0.10 – 0.05; asymptotic p-value = 0.52). In contrast, a significant positive correlation was observed between the selective constraints and the overall gene flux rate (Spearman’s R = 0.21, 95% CI: 0.14 – 0.27; asymptotic p-value = 3.3 ∗ 10 −9 ) ( Figure 4A ). Thus, genomes that experience intensive gene flux are subject to significantly stronger selective constraints than more stable genomes. Notably, the gene flux itself did not significantly correlate with the genome size (Spearman’s R = -0.05, 95% CI: -0.12 – 0.02; asymptotic p-value = 0.15), indicating that differences in genome sizes between ATGCs are not determined by a simple imbalance between gene acquisition and loss. When mapped to the prokaryote phylogeny, the correlation between the selective constraint and flux appeared to be more uniformly distributed across clades than the correlation between constraint and genome size ( Figure 4B ). Download figure Open in new tab Figure 4. Relationship between gene flux rate and ATGC-specific constraints across prokaryotes. (A) Scatterplot of gene flux rate (log scale) versus ATGC-specific constraint, where each point corresponds to a bacterial or archaeal ATGC (B) Distribution of Spearman’s correlation coefficients between flux rate and ATGC-specific constraint across the prokaryotic phylogeny. Major bacterial phyla containing more than 15 ATGCs, as well as Archaea, are highlighted with distinct colors. Internal nodes are colored according to the correlation among their descendant lineages; only nodes with more than 15 descendants are directly colored, whereas others inherit the color of their parent node. Discussion Previously, the evolution of prokaryote genomes was primarily considered to be driven either by selection for genome streamlining in large populations, as dictated by straightforward population genetics theory, or by selection for functional diversification as suggested by the observations on the positive correlation between selection strength and genomes size 1 , 5 , 16 , 19 – 22 . These aggregate approaches were necessitated in part by the lack of a sufficient number of groups of closely related genomes, for which selection could be accurately estimated, and in part by the relatively crude methodology that estimated selection from pairwise genome comparisons. Here we extended and refined the analysis of genome-wide selection regimes in prokaryotes by exploring ATGCs representing a broad diversity of bacteria and archaea, applying a more accurate, phylogenetic approach for dN/dS estimation, and decomposing selective constraints into gene-specific and lineage-specific components. This analysis validated the previous bulk estimates but showed that such estimates have been masking at least three distinct regimes of prokaryote evolution. The most striking observation is the dichotomy between the evolutionary regimes of the two domains of prokaryotes, bacteria and archaea. Whereas most bacterial taxa showed a positive correlation between selective constraints and genome size that implies selection for diversification, defining the overall correlation value, in archaea, a strong negative correlation was observed, suggestive of genome streamlining as the dominant selective factor. The observed negative association between selection constraint and genome size in Archaea is in an agreement with a previous report 29 , and reinforces the idea that archaeal genome evolution follows a distinct selective regime compared to Bacteria . Although there are many fewer archaeal ATGCs than bacterial ones, the difference between the selection regimes cannot be explained by a sampling error because extensive resampling of bacterial data did not reveal any subsets with nearly as strong streamlining signal as that observed for archaea. However, apart from the pronounced differences between bacteria and archaea, substantial heterogeneity was observed within the bacterial domain, with some groups showing no significant dependency of selection on genome size, and one order exhibiting the signature of streamlining. The causes of the pronounced dichotomy between the selective regimes of Archaea and Bacteria , and more generally, the heterogeneity of the selection landscapes among prokaryotes, remain to be explored and elucidated. Because selection constraints are inferred from the respective dN/dS values, they potentially can be driven by multiple underlying factors including effective population size, HGT rate, environmental stability and dynamics, and metabolic adaptation 30 , 31 . Thus, a plausible explanation of the heterogeneity of the evolutionary landscape could be that, for generalists, primarily, bacteria, that occupy ecologically complex niches and have access to large gene pools and diverse interaction partners, the defining selective factor is functional diversification via acquisition of multiple genes through HGT 32 , 33 . In contrast, for specialists, such as most archaea and some groups of bacteria, that typically inhabit stable, resource-limited or extreme environments, where high N e and limited ecological opportunities reduce the benefits of genetic innovation, streamlining appears to be the optimal evolutionary strategy 34 , 35 . The interplay between environmental pressures and population genetics parameters likely amplifies differences in selection regimes over long timescales. Compatible with the generalists vs specialists perspective, we observed a significant positive correlation between ATGC-specific selective constraints and the gene flux rate, but not with gene gain rate or loss rate separately, indicating that selection operates on the overall balance of opposing genome dynamics events, and is particularly strong in organisms with highly dynamic genomes. Gene acquisitions and deletions likely operate in a dynamic equilibrium that is shaped by both adaptive and neutral processes 31 , 36 . Consequently, for lineages that evolve under selection for diversification, intense gene flux provides a constant source of potentially beneficial genes, some of which are then integrated into molecular networks and start evolving under purifying selection. Although the positive correlation between genome size and selective constraints that is observed in most bacteria implies that many accessory genes are functionally relevant and evolve under strong purifying selection, this relationship is non-monotonic. Beyond an upper-intermediate genome size, the increase in selective constraint plateaus and can even decline ( Figure 1A , Figure 3 ), suggesting limits to the adaptive value of an expanding gene repertoire. This nonlinearity implies diminishing returns for adding genes after a certain point, due to factors such as functional redundancy, increased metabolic burden and other costs that selection cannot offset, leading to weaker constraint 37 . Furthermore, integration of new genes into the increasingly complex regulatory and metabolic networks require substantial molecular coordination, which may not be feasible beyond a certain genome size 38 , 39 , making unlikely the fixation of acquired genes. The lineage-specific selective constraints are highly consistent across functional classes of genes, scaling similarly with genome size. This covariation implies that, despite the broad variance of gene evolution rates, in most prokaryotes, changes in selection strength occur predominantly at a system-wide level, in accord with the concept of the universal pacemaker of genome evolution 40 . The genes associated with the mobilome are the only prominent exception, showing a substantially weaker correlation with genome-wide constraints. This decoupling of the mobilome from core cellular processes reflects the selfishness of mobile elements that propagate relying on their own molecular mechanisms rather than directly contributing to the overall fitness of prokaryotes 41 . Overall, our present work shows that genome size evolution in prokaryotes reflects a continuum of lineage-specific selective regimes rather than a single, universal evolutionary trend. The interplay between ecological conditions, population genetics, and molecular mechanisms can result in diverse evolutionary trajectories, highlighting the complexity and adaptability of prokaryotes. However, more theoretical work on models of evolution along with empirical analysis of genome evolution in diverse groups of bacteria and archaea are required to elucidate the underlying mechanisms of evolutionary processes that shape the composition of prokaryote genomes. Methods Genomic dataset and dN/dS calculation A total of 929 ATGCs were extracted from the ATGC database. Each genome was annotated by performing a competitive PSI-BLAST search of COG profiles 28 . When multiple orthologous groups within the same ATGCs mapped to the COG, the group containing the largest number of sequences was retained. The number of individual mutations in each COG was estimated by first reconstructing a phylogenetic tree from codon alignments using FastTree (gamma + GTR model) 42 , 43 , and then multiplying the total tree length by number of sites in corresponding alignment. Only alignments with at least 32 mutations were retained for downstream analyses. dN/dS values were calculated using the IQ-Tree with ModelFinder 27 , 44 , restricting search to models incorporating gamma distribution to account for variability in substitution rates among sites. A total of 2,724 alignments were simulated using AliSim 45 , employing empirical molecular values from 2,724 COGs of ATGC0001, including their observed dN/dS and transition/transversion ratios. Simulations were based on the phylogenetic tree of ATGC0001 that was pruned to match the taxa present in empirical sequences with 5,000 codons per alignment. Subsequently, 5,000 codons were trimmed to match the length of the corresponding real alignments. Identification of splittable ATGCs The phylogenetic tree of each ATGC was examined to identify branches that could split the tree into two subtrees, each with a phylogenetic depth greater than 0.2 and containing at least 20 genomes. If multiple branches met these criteria, the one that maximized the number of genomes on both subtrees was selected. During the analyses of 32 retained ATGCs, subtrees from 2 ATGCs produced largely unreliable dN/dS values (more than 30% equal to 1) and were excluded, leaving 30 ATGCs for subsequent analyses. Decomposition of dN/dS values into lineage-specific and gene-specific components and correlation analyses Lineage- and gene-specific components of dN/dS were estimated using an iterative decomposition approach. For a COG G in ATGC L , the model for its observed log-transformed dN/dS ratio ( r G , L ): where F L and R G are lineage- and COG-specific constraints respectively, and e G , L are residuals. Constraints were initialized at zero and were updated iteratively until convergence of the sum of squared residuals. Outliers with unlikely residuals under a normal approximation were removed, and the decomposition was repeated to obtain final estimates of constraints and residuals. Confidence intervals between ATGC-specific constraints and other metrics were estimated by performing 1,000 bootstrap replicates, followed by Bias-Corrected and Accelerated bootstrap (BCa), to account for bootstraps distribution and skewness 46 . Estimation of gene flux rate The data on the representation of ATGC genomes in COGs were converted into presence-absence binary sequences in FASTA format. Rooted ATGC trees together with the binary FASTA files were used as input for locally installed GLOOME software 47 . Reconstructed ancestral states were mapped onto the corresponding phylogenetic tree. For each COG, the tree was traversed from tips to root, and the change along every branch was calculated by subtracting the ancestral state from the descendant state. Positive differences were counted toward gains, and negative differences were counted toward losses. Changes across all COGs were then summed separately for gains and losses and normalized by the total tree length to obtain gain and loss rates. The overall flux rate is the ratio of gain rate to loss rate. Data Availability ATGC data that were used in this study is publicly available and can be found at https://ftp.ncbi.nih.gov/pub/wolf/ATGC/atgc-25/ . Analyzed genome IDs and derived data, including dN/dS values, splittable ATGCs, decomposition results, simulated alignments, and decomposition script are deposited in Zenodo under https://doi.org/10.5281/zenodo.17575555 . Funder Information Declared National Institutes of Health, https://ror.org/01cwqze88 , Intramural Research Program Footnotes https://doi.org/10.5281/zenodo.17575555 References 1. ↵ Lynch , M. & Conery , J.S . The origins of genome complexity . Science 302 , 1401 – 4 ( 2003 ). OpenUrl Abstract / FREE Full Text 2. ↵ Kuo , C.H. & Ochman , H . Deletional bias across the three domains of life . Genome Biol Evol 1 , 145 – 52 ( 2009 ). OpenUrl CrossRef PubMed 3. ↵ Mira , A. , Ochman , H. & Moran , N.A . Deletional bias and the evolution of bacterial genomes . Trends Genet 17 , 589 – 96 ( 2001 ). OpenUrl CrossRef PubMed Web of Science 4. ↵ Giovannoni , S.J. et al. Genome streamlining in a cosmopolitan oceanic bacterium . Science 309 , 1242 – 5 ( 2005 ). OpenUrl Abstract / FREE Full Text 5. ↵ Lynch , M . Streamlining and simplification of microbial genome architecture . Annu Rev Microbiol 60 , 327 – 49 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 6. ↵ Lopez-Perez , M. , Haro-Moreno , J.M. , Coutinho , F.H. , Martinez-Garcia , M. & Rodriguez-Valera , F. The Evolutionary Success of the Marine Bacterium SAR11 Analyzed through a Metagenomic Perspective . mSystems 5 ( 2020 ). 7. ↵ Wernegreen , J.J . Genome evolution in bacterial endosymbionts of insects . Nat Rev Genet 3 , 850 – 61 ( 2002 ). OpenUrl CrossRef PubMed Web of Science 8. ↵ Volland , J.M. et al. A centimeter-long bacterium with DNA contained in metabolically active, membrane-bound organelles . Science 376 , 1453 – 1458 ( 2022 ). OpenUrl CrossRef PubMed 9. Nakabachi , A. et al. The 160-kilobase genome of the bacterial endosymbiont Carsonella . Science 314 , 267 ( 2006 ). 10. Moran , N.A. & Bennett , G.M . The tiniest tiny genomes . Annu Rev Microbiol 68 , 195 – 215 ( 2014 ). OpenUrl CrossRef PubMed 11. ↵ Han , K. et al. Extraordinary expansion of a Sorangium cellulosum genome from an alkaline milieu . Sci Rep 3 , 2101 ( 2013 ). OpenUrl CrossRef PubMed 12. ↵ Shigenobu , S. , Watanabe , H. , Hattori , M. , Sakaki , Y. & Ishikawa , H . Genome sequence of the endocellular bacterial symbiont of aphids Buchnera sp. APS . Nature 407 , 81 – 6 ( 2000 ). OpenUrl CrossRef PubMed Web of Science 13. ↵ Andersson , S.G. et al. The genome sequence of Rickettsia prowazekii and the origin of mitochondria . Nature 396 , 133 – 40 ( 1998 ). OpenUrl CrossRef PubMed Web of Science 14. ↵ McCutcheon , J.P. & Moran , N.A . Extreme genome reduction in symbiotic bacteria . Nat Rev Microbiol 10 , 13 – 26 ( 2011 ). OpenUrl CrossRef PubMed 15. ↵ Moran , N.A . Microbial minimalism: genome reduction in bacterial pathogens . Cell 108 , 583 – 6 ( 2002 ). OpenUrl CrossRef PubMed Web of Science 16. ↵ Novichkov , P.S. , Wolf , Y.I. , Dubchak , I. & Koonin , E.V . Trends in prokaryotic evolution revealed by comparison of closely related bacterial and archaeal genomes . J Bacteriol 191 , 65 – 73 ( 2009 ). OpenUrl Abstract / FREE Full Text 17. ↵ Miyata , T. & Yasunaga , T . Molecular evolution of mRNA: a method for estimating evolutionary rates of synonymous and amino acid substitutions from homologous nucleotide sequences and its application . J Mol Evol 16 , 23 – 36 ( 1980 ). OpenUrl CrossRef PubMed Web of Science 18. ↵ Kryazhimskiy , S. & Plotkin , J.B . The population genetics of dN/dS . PLoS Genet 4 , e1000304 ( 2008 ). OpenUrl CrossRef PubMed 19. ↵ Bobay , L.M. & Ochman , H . Factors driving effective population size and pan-genome evolution in bacteria . BMC Evol Biol 18 , 153 ( 2018 ). OpenUrl CrossRef PubMed 20. Kuo , C.H. , Moran , N.A. & Ochman , H. The consequences of genetic drift for bacterial genome complexity . Genome Res 19 , 1450 – 4 ( 2009 ). OpenUrl Abstract / FREE Full Text 21. ↵ Sela , I. , Wolf , Y.I. & Koonin , E.V . Theory of prokaryotic genome evolution . Proc Natl Acad Sci U S A 113 , 11399 – 11407 ( 2016 ). OpenUrl Abstract / FREE Full Text 22. ↵ Martinez-Gutierrez , C.A. & Aylward , F.O . Genome size distributions in bacteria and archaea are strongly linked to evolutionary history at broad phylogenetic scales . PLoS Genet 18 , e1010220 ( 2022 ). OpenUrl CrossRef PubMed 23. ↵ Kristensen , D.M. , Wolf , Y.I. & Koonin , E.V . ATGC database and ATGC-COGs: an updated resource for micro- and macro-evolutionary studies of prokaryotic genomes and protein family annotation . Nucleic Acids Res 45 , D210 – D218 ( 2017 ). OpenUrl CrossRef PubMed 24. ↵ Yang , Z. & Nielsen , R . Estimating synonymous and nonsynonymous substitution rates under realistic evolutionary models . Mol Biol Evol 17 , 32 – 43 ( 2000 ). OpenUrl CrossRef PubMed Web of Science 25. Markova-Raina , P. & Petrov , D . High sensitivity to aligner and high rate of false positives in the estimates of positive selection in the 12 Drosophila genomes . Genome Res 21 , 863 – 74 ( 2011 ). OpenUrl Abstract / FREE Full Text 26. ↵ Anisimova , M. , Bielawski , J.P. & Yang , Z . Accuracy and power of the likelihood ratio test in detecting adaptive molecular evolution . Mol Biol Evol 18 , 1585 – 92 ( 2001 ). OpenUrl CrossRef PubMed Web of Science 27. ↵ Minh , B.Q. et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era . Mol Biol Evol 37 , 1530 – 1534 ( 2020 ). OpenUrl CrossRef PubMed 28. ↵ Galperin , M.Y. et al. COG database update 2024 . Nucleic Acids Res 53 , D356 – D363 ( 2025 ). OpenUrl CrossRef PubMed 29. ↵ Lyu , Z. , Li , Z.G. , He , F. & Zhang , Z. An Important Role for Purifying Selection in Archaeal Genome Evolution . mSystems 2 ( 2017 ). 30. ↵ Lynch , M . The frailty of adaptive hypotheses for the origins of organismal complexity . Proc Natl Acad Sci U S A 104 Suppl 1 , 8597 – 604 ( 2007 ). OpenUrl Abstract / FREE Full Text 31. ↵ Puigbo , P. , Lobkovsky , A.E. , Kristensen , D.M. , Wolf , Y.I. & Koonin , E.V . Genomes in turmoil: quantification of genome dynamics in prokaryote supergenomes . BMC Biol 12 , 66 ( 2014 ). OpenUrl CrossRef PubMed 32. ↵ McInerney , J.O . Prokaryotic Pangenomes Act as Evolving Ecosystems . Mol Biol Evol 40 ( 2023 ). 33. ↵ van Dijk , B. , Hogeweg , P. , Doekes , H.M. & Takeuchi , N . Slightly beneficial genes are retained by bacteria evolving DNA uptake despite selfish elements . Elife 9 ( 2020 ). 34. ↵ Giovannoni , S.J. , Cameron Thrash , J. & Temperton , B . Implications of streamlining theory for microbial ecology . ISME J 8 , 1553 – 65 ( 2014 ). OpenUrl CrossRef PubMed 35. ↵ Rodriguez-Gijon , A. et al. A Genomic Perspective Across Earth’s Microbiomes Reveals That Genome Size in Archaea and Bacteria Is Linked to Ecosystem Type and Trophic Strategy . Front Microbiol 12 , 761869 ( 2021 ). OpenUrl CrossRef PubMed 36. ↵ Jain , R. , Rivera , M.C. & Lake , J.A . Horizontal gene transfer among genomes: the complexity hypothesis . Proc Natl Acad Sci U S A 96 , 3801 – 6 ( 1999 ). OpenUrl Abstract / FREE Full Text 37. ↵ Koonin , E.V. & Wolf , Y.I . Constraints and plasticity in genome and molecular-phenome evolution . Nat Rev Genet 11 , 487 – 98 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 38. ↵ Croft , L.J. , Lercher , M.J. , Gagen , M.J. & Mattick , J.S . Is prokaryotic complexity limited by accelerated growth inregulatory overhead? Genome Biology 5 , P2 ( 2003 ). OpenUrl CrossRef 39. ↵ Lercher , M.J. & Pal , C . Integration of horizontally transferred genes into regulatory interaction networks takes many million years . Mol Biol Evol 25 , 559 – 67 ( 2008 ). OpenUrl CrossRef PubMed Web of Science 40. ↵ Snir , S. , Wolf , Y.I. & Koonin , E.V . Universal pacemaker of genome evolution . PLoS Comput Biol 8 , e1002785 ( 2012 ). OpenUrl CrossRef PubMed 41. ↵ Werren , J.H . Selfish genetic elements, genetic conflict, and evolutionary innovation . Proc Natl Acad Sci U S A 108 Suppl 2 , 10863 – 70 ( 2011 ). OpenUrl Abstract / FREE Full Text 42. ↵ Yang , Z . Maximum likelihood phylogenetic estimation from DNA sequences with variable rates over sites: approximate methods . J Mol Evol 39 , 306 – 14 ( 1994 ). OpenUrl CrossRef PubMed Web of Science 43. ↵ Price , M.N. , Dehal , P.S. & Arkin , A.P . FastTree 2--approximately maximum-likelihood trees for large alignments . PLoS One 5 , e9490 ( 2010 ). OpenUrl CrossRef PubMed 44. ↵ Kalyaanamoorthy , S. , Minh , B.Q. , Wong , T.K.F. , von Haeseler , A. & Jermiin , L.S . ModelFinder: fast model selection for accurate phylogenetic estimates . Nat Methods 14 , 587 – 589 ( 2017 ). OpenUrl CrossRef PubMed 45. ↵ Ly-Trong , N. , Naser-Khdour , S. , Lanfear , R. & Minh , B.Q . AliSim: A Fast and Versatile Phylogenetic Sequence Simulator for the Genomic Era . Mol Biol Evol 39 ( 2022 ). 46. ↵ Efron , B . Better Bootstrap Confidence Intervals . Journal of the American Statistical Association 82 , 171 – 185 ( 1987 ). OpenUrl CrossRef Web of Science 47. ↵ Cohen , O. , Ashkenazy , H. , Belinky , F. , Huchon , D. & Pupko , T . GLOOME: gain loss mapping engine . Bioinformatics 26 , 2914 – 5 ( 2010 ). OpenUrl CrossRef PubMed Web of Science View the discussion thread. Back to top Previous Next Posted November 27, 2025. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following The heterogeneous selection landscape of genome evolution in prokaryotes Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share The heterogeneous selection landscape of genome evolution in prokaryotes Roman Kogay , Svetlana Karamycheva , Nash D. Rochman , Yuri I. Wolf , Eugene V. Koonin bioRxiv 2025.11.26.690804; doi: https://doi.org/10.1101/2025.11.26.690804 Share This Article: Copy Citation Tools The heterogeneous selection landscape of genome evolution in prokaryotes Roman Kogay , Svetlana Karamycheva , Nash D. Rochman , Yuri I. Wolf , Eugene V. Koonin bioRxiv 2025.11.26.690804; doi: https://doi.org/10.1101/2025.11.26.690804 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Evolutionary Biology Subject Areas All Articles Animal Behavior and Cognition (7637) Biochemistry (17705) Bioengineering (13899) Bioinformatics (41970) Biophysics (21463) Cancer Biology (18605) Cell Biology (25526) Clinical Trials (138) Developmental Biology (13385) Ecology (19911) Epidemiology (2067) Evolutionary Biology (24329) Genetics (15615) Genomics (22514) Immunology (17743) Microbiology (40424) Molecular Biology (17194) Neuroscience (88650) Paleontology (667) Pathology (2835) Pharmacology and Toxicology (4827) Physiology (7648) Plant Biology (15160) Scientific Communication and Education (2046) Synthetic Biology (4302) Systems Biology (9825) Zoology (2271)